Background: Serum retinol binding protein (RBP4) concentrations are elevated in insulin resistant states such as obesity, the metabolic syndrome, and type 2 diabetes, and RBP4 plays a causal role in insulin resistance and glucose intolerance. RBP4 is the sole specific transport protein for Vitamin A (retinol) in blood. While liver is the primary source of circulating RBP4, adipocytes may be an important secondary source in insulin resistant states. Despite significant clinical interest in RBP4 as a therapeutic target, tissue sources of circulating RBP4 and the basic mechanisms by which RBP4 interacts with cells to cause insulin resistance remain unclear. Based on tissue mRNA content, it has been inferred that up to 20% of circulating RBP4 is normally produced by adipocytes. In insulin resistant states, adipocyte mRNA content is dramatically increased and highly correlated with serum RBP4 concentrations, whereas hepatocyte RBP4 mRNA does not change. Based on these observations it has been suggested that increased expression of RBP4 in adipocytes is the sole source of increased circulating RBP4 in insulin resistant states, but this has never been formally tested. The tissue source of RBP4 in insulin resistant states is important because adipocytes do not store retinol efficiently and may secrete a distinct form of RBP4 that is not bound to retinol, conventionally referred to as apo-RBP4. Hepatocytes, on the other hand, are known to primarily secrete retinol-bound RBP4, which typically comprises 80% of circulating RBP4 and is referred to as holo-RBP4. It has recently been reported that concentrations of apo-RBP4 are increased 2- to 3-fold in insulin resistant human subjects, and our preliminary data demonstrate a similar increase of apo-RBP4 in insulin resistant obese mice. Because apo-RBP4 is defined simply as the fraction of RBP4 not bound to retinol, it remains possible that apo-RBP4 binds and transports one or more hydrophobic ligands (other than retinol) that mediate insulin resistance upon their delivery to tissues. Consistent with this, our preliminary data indicate that even very high amounts of dietary retinol do not cause insulin resistance in mice, and furthermore, treatment of mice with retinoic acid, a metabolite of retinol and broad regulator of gene transcription, causes leanness and enhances insulin sensitivity in mice. Together these data support the possibility that RBP4 and its potential non-retinol ligand(s) produced by adipocytes are materially distinct from the retinol-bound holo-RBP4 produced by hepatocytes. Hypothesis: Insulin resistance associated with elevated serum RBP4 is principally caused by the fraction of RBP4 secreted from adipocytes. Specific aims: (i) To test whether apo-RBP4 produced by adipocytes contributes to the elevation of RBP4 and insulin resistance associated with obesity; (ii) To test whether holo-RBP4 produced by hepatocytes plays a protective role in maintaining insulin- glucose homeostasis; (iii) To test whether circulating apo-RBP4 and adipocyte-secreted RBP4 transport one or more non-retinol ligands that cause insulin resistance at the cellular level. Methods: We will employ mouse models of tissue specific-knockout of RBP4 expression in hepatocytes or adipocytes generated by means of Cre/loxP technology to study how RBP4 produced in each of these tissues contributes to systemic insulin resistance and glucose intolerance under normal conditions and in the setting of high fat diet. We will further employ a novel technique to produce recombinant RBP4 in vivo under insulin resistant conditions, in order to test its capacity for inducing insulin resistance at the cellular level.